Three-dimension (3D) graphics applications typically start a frame by clearing the framebuffer. The clear operation can be very time consuming and wasteful of system resource such as memory operations since the clear operation is expected to touch every pixel in the framebuffer. GPU implementations optimize the clear operation by delaying the activity of touching any pixel by tracking what pixels have been over-written and then performing the clear operation only on the pixels that were not touched during rendering. However, certain pixels which were cleared in a previous frame and have not been written to since they were cleared are cleared once more even with such delayed clear optimizations.